Lash up Direct Conversion Receiver

[Skywave]

Lawrence: harmonic generator, Racal RA-17: scanned copy of hand-drawn sketch is below.

[attachment=4782]

To fully appreciate what the HG does in the RA-17 and why it's needed, a brief description how it works is required. In essence, the L/C filter (after the HG valve itself) is a low-pass filter: passes all freqs. up to 32 MHz. The O/P from that filter is fed to a mixer. The other input to that mixer is the 1st. VFO (MHz tuning). The output of that mixer is a 37.5 MHz drive signal that is fed through a 37.5 MHz BPF to a second mixer. Since the 37.5 MHz BPF is +/- 150 kHz wide, only the appropriate harmonic from the HG LPF produces 37.5 MHz drive to that other mixer. (It's that 37.5 MHz BPF that 'calls the shots' here). The consequence of that is that the 1st. VFO must be tuned in 1 MHz steps. to produce that drive signal. (The reason as to why a 37.5 MHz drive signal is required in the first place is not discussed here: it's a separate issue: but please feel free to ask!).

All that is made clear by a few numerical examples:
1st. VFO tunes from 40.5 to 69.5 MHz. To produce a 37.5 MHz IF drive signal from the mixer and its 37.5 MHz (+/- 150 kHz) BPF combination, the appropriate harmonic from the HG will be as follows.

1st. VFO:    40.5  41.5  42.5  43.5  44.5 . . . . . 69.5 MHz
Harmonic:    3.0    4.0    5.0   6.0    7.0 . . . . . 32.0 MHz
37.5 MHz:    37.5 . . . . . . . . . . . . . 37.5         37.5 MHz

Al.

[pwdrive]

Hi Al, thanks for that, the circuit you enclosed is the same as the one I have been looking at, so as I see it the 1 mhz xtal osc feeds into the harmonic generator the output of which is passed through the 32 mhz LPF and then to the mixer.
So am I right also in thinking that at the output from the 32 mhz LPF contains ALL the harmonics of the 1 mhz xtal osc eg: 2,3,4,5,6,7 mhz etc upto 32 mhz and thence after roll off suppresses harmonics above 32 mhz and beyond.
In your judgement could you let me know.
Please note a full Wadley loop is not in mind! just interested to see if I can get all the harmonics of a 1 mhz xtal osc upto around 30 mhz for use as local osc in the regenerodyne based receiver I am thinking of.

Lawrence

[Skywave]

So am I right also in thinking that at the output from the 32 MHz LPF contains ALL the harmonics of the 1 MHz xtal osc e.g.: 2, 3, 4, 5, 6, 7 MHz etc up to 32 MHz and thence, after roll off, suppresses harmonics above 32 MHz and beyond?
Lawrence

You got it: yes.

Al.

[pwdrive]

Thanks Al, in theory then I should be able to use that idea as a band set osc for the regenerodyne, the regen detector tuning 500khz each way over the conversion frequency ie if conversion freq. is 3 mhz i can pick out the appropriate harmonic from the 1 mhz xtal osc/harmonic generator with a tuned circuit (probably need 2 or 3, switchable, to cover up to 30 odd mhz) and inject that into the mixer and tune 500 khz either side of the difference freq. (the IF) eg: 10th harmonic = 10 mhz, 10 mhz minus IF of 3 mhz = 7 mhz therefore regen det can tune in effect 6.5 mhz to 7.5 mhz using the 10th harmonic (or the 4th) from the xtal osc/harmonic generator using it as an osc feed to the mixer.... and so on in 1 mhz steps.
Am I nuts or what...I am pretty sure I can build a nice smooth consistant regen det at 3 mhz (higher freq. might be a problem) so if I can crack the harmonic generator idea that would be good, if i cannot get that idea to work properly I will incorporate a buffer amp into the receiver and use the sig. gen. to feed the buffer amp and use that as a bandset osc rather than going the switched xtal/doubler route as I would like to cover the whole SW band.

Lawrence.

[pwdrive]

Just as an addition, I suppose I could test the theory of the 1 mhz xtal harmonic generator out by knocking up a class C job and feeding the sig. gen. in and see what I can tune out of the anode tank, trouble is no 'scope so might be guess work and hunch. If it fails I will post results anyway, unless someone who has a 'scope wants to lash one up to compare notes...no transistor(s) just valve(s)

Lawrence

[Skywave]

The critical bit of the HG in the Racal is, of course, the 32 MHz LPF. I don't know comfortable you are with the design procedure for such an L/C filter and even if you manage to produce a theoretical design on paper, there remain the three problems of (a) winding the coils, (b) measuring their inductance and © testing the performance of the completed filter. Appropriate test equipment is clearly called for: inductance meter, signal generator (ideally a sweeper) and a 'scope for a start.

Al.

[pwdrive]

Thanks Al I thought that might be the case (designing/building the LPF) but tell me is the filter primarily for suppressing harmonics above 32 mhz and if so does it have any real effect in producing harmonics below that said frequency eg do I need it to get sub 32 mhz harmonics.

Lawrence

[pwdrive]

The Sun is slowly going down so the 7 mhz reception (40 meter band) is getting very good now, plenty of amature transmissions on the present lash up, it has been working most of the day, as far as I can ascertain the regen det. has not drifted to much since the initial warm up and no discernable breakthrough as yet from the strong AM transmissions that are on the top edge of that amature band. I have made a simple mod. to the AF choke by tuning it with fixed capacitor to make a crude AFpass filter which has helped reduce the hash somewhat, I can only guess at this filters bandwidth!
So quite happy with it so far.

Lawrence.

[Skywave]

Lawrence: the 32 MHz H.G. filter. (Ref. your post #27)

To my understanding, in essence it's just like any other LPF: ideally, flat response to 32 MHz then a roll off to full attenuation. The highest freq. required from the filter is 32 MHz. The next 'available' freq. is 33 MHz, which isn't 'far away', and that must be suppressed: I would expect something like -20 dB at 33 MHz. I doubt if insertion loss is critical - which gives some latitude to the coil design - provided that the flatness of the pass-band (for the freqs. 1 MHz to 32 MHz) is fairly even (say +/- 2 dB), so that the output harmonics of the 1 MHz input are at all about the same level.
So overall, a LPF with quite a demanding performance: stop band -20 dB within 1 MHz (32 - 33 MHz) and a 'flat-ish' pass-band. Hence several L/C sections required; R/C sections won't do.

Al.

[pwdrive]

Thanks for that Al, it is more or less what I thought but needed a second opinion. There are a few multi element filter design programs (L/C) on the net, the problem is of course what you said earlier as regarding home brewing one. I think Mini Circuits (Of SBL1 mixer fame) do LPF's for reasonable money but the roll off seams to start a lot higher than 32 mhz from what I remember plus I am not sure how they would match up impedance wise for valves, I will try the cave man approach first as in sig gen into class C amp and try and tune an anode tank circuit to extract harmonics...A bit of experimenting.
By the way how did your plumbing job go?

Lawrence.